Specialized Custom DOE for Experienced Experimenters

Presenter: Tom Donnelly

Augmenting Designs to Add Constraints and Support Second Order Models

See how to:

  • Improve a cooking experimental design by adding additional runs that exclude extreme settings known to be deleterious to the process (2:52)
    • Add linear constraints and group new runs into blocks to make it easier to examine design
    • Use Graph Builder to visualize location of new trials in the design space
  • Model a process to convert captured carbon dioxide emissions into plastic (9:59)
    • Augment design to support second order model suitable when more than a few factors are significant
    • Use Restrict option to indicate than a term cannot enter the model unless all the terms above it in the hierarchy have been entered 
    • Use Stopping Rule Stopping Rule to determine if corrected Akaike's Information Criterion (AICc) or Bayesian Information Criterion (BIC) will be used to select best model
    • Understand why and how to use Square Root transformation
    • Compare original and new models, and then choose best model
  • Create and then augment carbon dioxide emissions design from 24 to 36 runs (19:00)
    • Create and interpret design table

Q&A included at times 6:11, 6:59, 21:22, 22:31, 23:35, 25:05 and 26:54.


Back to Top